This article challenges the ‘‘diversity debit” hypothesis that ethnic diversity has a negative impact on social, economic, and political outcomes. It argues that rather than focus on why this is the case, research should seek to explore the underlying mechanisms of such negative relationship, specifically why these relationships may differ across sub-national budgetary and welfare outcomes. It argues that ethnic fractionalisation is not associated with the under-provision of public goods and does, in fact, have a positive relationship with some key welfare outcomes. The paper considers key explanatory hypotheses against the Zambian data to identify promising areas for such theory development.
The article builds on the available literature, and draws on new district-level analysis which uses a new dataset from administrative, budget, and survey data,. This data covers a broader range of public goods outcomes than previous work, including information on both budgetary and welfare outcomes.
Results most directly counter the diversity debit hypothesis – there is a statistically significant positive relationship found between ethnic diversity and welfare outcomes in this case.This is robust across a variety of alternative specifications and models:
Ethnic diversity and budgetary outcomes:
- The results find a clearly negative relationship between ethno-linguistic fractionalisation and central government expenditure at district level in both sectors and across budget lines.
- The log of district population and the distance by road to Lusaka appear to be good predictors of differences in budget allocations between districts. The distance to Lusaka is significant and negative throughout for total and per capita expenditure, except for grants to basic schools and health service delivery, where it is significant only in some model specifications.
- The results do not provide particularly strong evidence for political targeting of recurrent health and education expenditure.
Ethnic diversity and welfare outcomes:
- The results show a clearly positive relationship across specifications between ethnic fractionalisation and primary school enrollment, but none with the other outcome variables.
- Results from the health sector are in line with those from the education sector. There is a positive effect of ethnic diversity on all immunization rates, the under-five mortality rates, and the share of underweight children under five. The only exception is the number of beds in health facilities per 1,000 population, for which we find a significant negative relationship.
The overall positive association could be consistent either with diversity driving improved outcomes, or improved outcomes driving higher diversity. However, a number of explanations, such as the effect of migration or optimal sorting, do not seem to explain the results. A combined understanding of these explanations do highlight why ethnic diversity does not necessarily undermine public goods provision and the possible channels that may underlie a diversity dividend.
Diversity can be good for communities, not only for normative reasons, but also because, under some conditions, it can support concrete welfare gains. Given the more nuanced empirical relationships now documented here, the key question for future work is not so much ‘‘why does ethnic diversity undermine public goods provision?” but ‘‘when does ethnic diversity support public goods provision and aggregate welfare and when does it undermine it?” and ‘‘why and under what conditions does a diversity divided exist?”. Further research is needed to explore this fully.